Usefulness of antikeratin immunoreactivity in osteosarcomas of the jaw Jin Kim, DDS,a Gary L. Ellis, DDS,b and Thomas A. Mounsdon, DDS,c Seoul, Korea, Washington, D.C., and Bethesda, Md.

YONSEI UNIVERSITY SCHOOL OF DENTISTRY, ARMED FORCES INSTITUTE OF PATHOLOGY, AND

NAVAL DENTAL SCHOOL

The immunohistochemical typing of cytoplasmic intermediate filaments has proved helpful to the pathologist in classifying poorly differentiated malignant neoplasms. In general, identification of keratin-type intermediate filaments has been associated with epithelial histodifferentiation, but several exceptions to this generalization have been reported in the literature. A recent report identified false-positive immunostaining for keratin in osteosarcomas of the jaws that was attributed to cross-reactivity induced by enzyme digestion of the tissue specimens before immunostaining. Because the jaws are unique in the skeletal system because of their relatively high incidence of intraosseous epithelial neoplasms, false-positive immunoreactions for keratin could complicate differentiating sarcomatoid epithelial neoplasms from poorly differentiated osteosarcomas. To evaluate this possible pitfall in our laboratory, eight osteosarcomas of the jaws were evaluated for keratin immunostaining with polyclonal and monoclonal antibodies on tissue sections that had been enzymatically treated with protease. No immunostaining was demonstrated in these tumors. Repudiation of the usefulness of antikeratin immunohistochemistry for intraosseous jaw tumors was not confirmed with the procedures used in our laboratory. (ORAL SURC ORAL MED ORAL PATHOL 1991;72:213-7)

I mmunohistochemistry, one of the important diag- nostic and research tools in surgical pathology, can be helpful in classifying malignant tumors. In particular, the typing of intermediate filaments that characterize the cytodifferentiation of cells may provide evidence of the his&differentiation of poorly differentiated malignant tumors. Indistinguishable by electron mi- croscopy, the intermediate filaments are composed of a related family of protein subunits that have been biochemically and immunologically separated into five major classes that have generally been shown to be cell and tissue differentiation specific.’ Keratin, desmin, vimentin, glial fibrillary acidic protein, and

The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Department of the Army, the Department of De- fense, or the Department of Veterans Affairs. *Division of Oral Pathology, Yonsei University School of Dentistry, Seoul. bDepartment of Oral Pathology and the Department of Veterans Affairs Special Reference Laboratory for Pathology, Armed Forces Institute of Pathology, Washington. CDepartment of Oral Pathology, Naval Dental School, Bethesda. 7/14/28234

neurofilaments are associated with epithelial, muscle, mesenchymal, astroglial, and neuronal cells, respec- tively. Gabbiani et a1.2 and Osborn and Weber3 have shown evidence that differentiation-related interme- diate filament specificity is preserved in neoplasia.

However, exceptions to these generalizations have been noted. Cytokeratins have been demonstrated in such apparently nonepithelial neoplasms as synovial sarcoma,4 rhabdomyosarcoma,5 astrocytoma,6 epi- thelioid sarcoma,7 leiomyosarcoma,8 and malignant fibrous histiocytoma.9 Cells positive for glial fibrillary acidic protein have been found outside the nervous system in pleomorphic adenomas of salivary gland. lo Additionally, the simultaneous expression of two dif- ferent types of intermediate filaments has been noted in some neoplastic cell lines.2, 5-7, lo-i9 The most fre- quently occurring exception has been the coexpres- sion of vimentin with one of the other types of inter- mediate filaments. Identification of tumor types that may react contrary to the generally accepted princi- ple of intermediate filament distribution has impor- tant implications for the evaluation of a differential diagnosis for poorly differentiated malignant neo- plasms.

No specific immunohistochemical markers have

213

214 Kim, Ellis, and Mounsdon ORAL SURG ORAL MED ORAL PATHOL August 199 I

yet been elucidated for osteosarcoma, the most com- mon malignant tumor of bone, although vimentin, HLA-DR, (~1 -antichymotrypsin, ~1 -antitrypsin, and soybean agglutinin lectin binding can be found nonspecifically. 20-22 Recently Regezi et a1.20 reported an immunohistochemical study of osteosarcomas of the jaw in which they demonstrated immunofeactiv- ity for keratin in 7 of 16 tumors. They attributed the immunostaining of osteosarcomas with antikeratin antibodies to cross-reactivity induced by trypsin di- gestion of the tissues and commented on this poten- tial diagnostic pitfall.

Primary and metastatic malignant epithelial neo- plasms occur within the jaws,23, 24 and sometimes these carcinomas may have a sarcomatoid appear- ance. Therefore false-positive immunoreactions for keratin antigens have serious implications for the evaluation of poorly differentiated malignant in- traosseous neoplasms that occur in these sites.

In an attempt to evaluate the false-positive results obtained by Regezi et a1.20 in our laboratory, we per- formed an immunohistochemical analysis with anti- keratin antibodies on a series of osteosarcomas of the jaw.

MATERIAL AND METHODS

Eight cases coded as osteosarcomas of the jaws were retrieved from the files of the Armed Forces In- stitute of Pathology. These cases represented material submitted by civilian and military pathologists for consultation. Required for inclusion in the study was the availability of paraffin-embedded tissue for im- munohistochemical evaluation. All tissues had been formalin fixed and decalcified. Hematoxylin-eosin- stained slides were reviewed and the diagnosis con- firmed in each case.

Both polyclonal and monoclonal antikeratin anti- bodies were used to demonstrate immunoreactivities. The polyclonal antibody (PAb) was rabbit antisera to keratin. The antikeratin PAb was affinity isolated as previously described.’ Briefly, antiserum was derived in rabbits with keratin proteins isolated from human plantar calluses according to the method of Sun and Green.25 The antibodies were affinity isolated with purified keratin proteins linked to cyanogen bromide- activated sepharose 4B (Pharmacia Chemical Co., Piscataway, N.J.). Monoclonal antibodies (MAbs) used in the study were AEl /AE3 (Boehringer-Mann- heim, Indianapolis, Ind.). The MAbs AE1/3 are a combination of two MAbs that recognize most acidic and basic keratin proteins.26 All the immunostaining was performed with predigestion with 0.05% protease type VIII (Sigma Chemical Co., St. Louis, MO.) in 0.1 mol/L phosphate buffer, pH 7.8, at 37” C for 7 minutes.27 We found that protease is more efficient in exposing immunodeterminants than trypsin. Immu-

nostaining was completed with modifications of the peroxidase-antiperoxidase procedure as previously described.’ For negative controls duplicate sections were immunostained as described previously, except normal rabbit serum (for PAb) or normal mouse se- rum (for MAbs) were substituted for the primary an- tibody. Mucosal epithelium was used as a positive control. Sections were developed in 0.016% diami- nobenzidine tetrahydrochloride (Sigma) with 0.05% hydrogen peroxide in phosphate-buffered saline solu- tion at room temperature for 15 minutes.

RESULTS

All but one of the eight osteosarcomas occurred in women. The patients’ ages ranged from 23 years to 79 years and were evenly distributed through these decades. Four tumors each affected the maxilla and the mandible.

The most frequent symptom was tumorous swelling of the affected sites (four cases). In two cases the ini- tial symptom was gingival bleeding and one of these was a recurrent tumor. One patient with sarcoma in the maxillary antrum had recurrent epistaxis. Pares- thesia of the lower lip was the chief complaint of one patient with osteosarcoma of the mandible.

Seven neoplasms were classified as osteoblastic- type osteosarcomas and were characterized by the proliferation of malignant osteoblasts with deposition of variable amounts of osteoid matrix. One of these tumors demonstrated well-formed bony trabeculae that simulated benign osteoblastoma. One case was categorized as a fibroblastic-type osteosarcoma that showed only focal areas of osteoid formation in a fi- brosarcomatous background.

Neither polyclonal nor monoclonal antikeratin an- tibodies reacted to the tumor cells in any case by im- munohistochemical analysis. Normal mucosal epithe- lium showed intense reactivity with both types of an- tibodies.

DISCUSSION

Osteosarcomas in the jaws differ clinically from those in the long bones by the higher peak age of oc- currence, the lower rate of metastasis, and the more favorable prognosis. 28-33 The histologic features of osteosarcoma of the jaws, however, are similar to those of the long bones and include parosteal, peri- osteal, osteoblastic, chondroblastic, fibroblastic, and telangiectatic types.34-39

Within the skeletal system, the jaws are unique be- cause of the relatively frequent occurrence of in- traosseous epithelial neoplasms. Most of these epi- thelial neoplasms are derived from odontogenic epi- thelium, but metastatic carcinomas and extensions from adjacent mucosal and salivary gland carcinomas also occur. In most instances the identification of a

Volume 72 Antikeratin immunoreactivity in osteosarcoma of jaw 2 15 Number 2

neoplasm as a carcinoma can be accomplished on a morphologic basis. However, the determination of the histodifferentiation of poorly differentiated neoplasms can be problematic.

With a few exceptions, immunocytochemistry for specific intermediate filament type has proved to be a useful tool for evaluating the histodifferentiation of malignant neoplasms. Immunoreactivity of tumor cells with antikeratin antibodies is generally accepted as indicative of epithelial differentiation. However, intelligent analysis of immunohistochemical studies requires an awareness not only of neoplasms that may react contrary to conventional wisdom but also of sit- uations that may produce false-positive reactions. The false-positive immunostaining with antikeratin anti- bodies in a significant percentage of osteosarcomas of the jaws, reported by Regezi et a1.,20 raised serious questions about the validity of using antikeratin im- munohistochemistry to evaluate epithelial differenti- ation of poorly differentiated malignant neoplasms in the jaws. In an attempt to evaluate the problem these authors had with false-positive results in our labora- tory, we analyzed a series of eight osteosarcomas of the jaws with polyclonal and monoclonal antikeratin antibodies. None of our cases showed positive immu- nostaining.

To eliminate false-positive immunohistologic stain- ing, methods including the dilution of primary an- tiserum and application of normal serum of the spe- cies from which bridging antibody was derived have been developed. 40-42 Fixation is a primary factor in the preservation of tissue antigenicity. According to some investigators43* 44 sublimate-formaldehyde fix- ative is the best for preservation of tissue antigenici- ty, but this fixative requires precautions because of its toxicity and corrosive property.43 Additionally, Banks45 reported that direct primary fixation with this agent resulted in increased nonspecific binding of antisera. Battifora and Kopinski46 recommended al- cohol fixation for keratin immunostaining. Bone tis- sue requires decalcifying agents as well as fixatives. Ethylenediaminetetraacetic acid, formic acid, acetic acid, and acetic acid-formaldehyde-saline solution were reported to preserve immunoreactivity we11,47-49 but hydrochloric acid- and nitric acid-containing agents resulted in variable stainability and poor pres- ervation of histologic structures.47 Pretreatment of tissue sections with proteolytic enzymes can enhance immunoreactivity by exposing antigenic sites that were masked during fixation.50-53 Enzyme concentra- tion, treatment time, and temperature are factors that influence results. According to the study of Mepham et a1.,50 trypsinization needs a longer tissue exposure period than pronase digestion.

Regezi et a1.20 ascribed the false-positive immu- nostaining in their study to cross-reactivity induced

by predigestion of the tissue sections with trypsin. Matthews and Mason47 reported that keratin was poorly detected in both decalcified and nondecalcified tissues because of the long period of trypsinization required for unmasking keratin antigens. In our lab- oratory we use protease digestion to enhance immu- noreactivity with certain antibodies, including anti- keratin antibodies. The protease we use has greater digestive activity than trypsin but lesser digestive ac- tivity than pronase. The difference in enzymes em- ployed may explain the different results obtained in the two studies.

MAbs can improve the specificity of immunoreac- tivity because they react with only one epitope of an antigen. However, it is possible that an MAb may cross-react with similar epitopes on an unrelated antigen. Our study employed MAbs AE1/3, which recognize a wide variety of acidic and basic kera- tins.54-56 Regezi et a1.20 applied MAbs 902 and 903. MAb 902 is derived from human hepatoma cells and reacts with 54 kd keratin type, and MAb 903 is de- rived from human stratum corneum and reacts with 49 5 1,57, and 66 kd keratins.57, 58 Although it is un- likely that keratin antigens reactive with MAbs 902/ 903 could escape detection with MAbs AE1/3, it is possible that MAbs 9021903 may react with an epitope on a nonkeratin antigen that is not recognized by MAbs AE1/3. In addition, PAb demonstrated no positive immunoreaction in our study.

Regezi et a1.20 reported that most of the antikera- tin immunostaining occurred in areas of chondroid differentiation. The tumors we analyzed were mostly osteoblastic, but two of three osteoblastic osteosarco- mas in the study of Regezi et a120 were immuno- stained with antikeratin antibodies. Loning et a12’ were also unable to detect other types of intermediate filaments except vimentin in osteosarcoma. The im- portant conclusion from this study is that false-posi- tive immunostaining of osteosarcomas of the jaw for keratin antigens was not repeatable in this study with the procedures used in our laboratory. Currently, re- pudiation of the usefulness of antikeratin immuno- histochemistry on poorly differentiated malignant neoplasms of the jaws has not been validated in our laboratory. Other laboratories may wish to repeat these studies to clarify this issue further.

REFERENCES Lazarides E. Intermediate filaments as mechanical integrators of cellular space. Nature 1980;283:249-56. Gabbiani G. Kaoanci Y. Barazzone P, Franke WW. Immuno- chemical identihcation of intermediate-sized filaments in hu- man neoplastic cells: a diagnostic aid for the surgical pathol- ogist. Am J Pathol 1981;104:206-16. Osborn M, Weber K. Intermediate filaments: cell-type-spe- cific markers in differentiation and pathology. Cell 1982; 3 I :303-6. Corson JM, Weiss LM, Banks-Schlegel SP, Pinkus GS. Ker-

216 Kim, Ellis, and Mounsdon

5.

6.

I.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

Weiss SW, Bratthauer GL, Morris PA. Post-irradiation ma-

atin proteins and carcinoembryonic antigen in synovial sarco- mas: an immunohistochemical study of 24 cases. Hum Path01

lignant fibrous histiocytoma expressing cytokeratin. Am J

1984;15:615-21. Miettinen M, Rapola J. lmmunohistochemical spectrum of

Surg Path01 1988;12:554-8.

rhabdomyosarcoma and rhabdomyosarcoma-like tumors: ex- pression of cytokeratin and the 68-kD neurofilament protein. Am J Surg Pathol 1989;13:120-32. Cosgrove M, Fitzgibbons PL, Sherrod A, Chandrasoma PT, Martin SE. lntermediate filament expression in astrocytic neoplasms. Am J Surg Pathol 1989;13:141-5. Chase DR, Enzinger FM, Weiss SW, Langloss JM. Keratin in epithelioid sarcoma: an immunohistochemical study. Am J Surg Pathol 1984;8:435-41. Norton AJ, Thomas JA, Isaacson PG. Cytokeratin-specific monoclonal antibodies are reactive with tumors of smooth muscle derivation: an immunocytochemical and biochemical study using antibodies to intermediate filament cytoskeletal proteins. Histopathology 1987;11:487-99.

Nakazato Y, lshizeki J, Takahashi K, Yamaguchi H, Kamei T, Mori T. Localization of S-100 protein and glial fibrillary acidic protein-related antigen in pleomorphic adenoma of the salivary glands. Lab Invest 1982;46:621-6. Caselitz J, Osborn M, Seifert G, Weber K. Intermediate-sized filament proteins (prekeratin, vimentin, desmin) in the normal parotid gland and parotid gland tumors. Virchows Arch [A] 1981;393:273-86. Krepler R, Denk H, Artlieb U, Mall R. Immunocytochemis- try of intermediate filament proteins present in pleomorphic adenomas of the human parotid gland: characterization of dif- ferent cell types in the same tumor. Differentiation 1982; 21:191-9. Caselitz J, Becker J, Seifert G, Weber K, Osborn M. Coex-

Ellis GL, Langloss JM, Enzinger FM. Coexpression of keratin

pression of keratin and vimentin filaments in adenoid cystic carcinomas of salivary glands. Virchows Arch [A] 1984;

and desmin in a carcinosarcoma involving the maxillary alve-

403:337-44. von Bassewitz DB, Roessner A, Grundmann E. Intermediate-

olar ridge. ORAL SURG ORAL MED ORAL PATHOL 1985;

sized filaments in cells of normal human colon mucosa, adenomasand carcinomas. Pathol Res Pratt 1982;175:238-55.

60:410-6.

Miettinen M, Franssila K, Lehto V-P, Paasivuo R, Virtanen I. Expression of intermediate filament proteins in thyroid gland and thyroid tumors. Lab Invest 1984;50:262-70. Denk H, Krepler R, Artlieb U, et al. Proteins of intermediate filaments: an immunohistochemical and biochemical approach to the classification of soft tissue tumors. Am J Path01 1983; 110:193-208.

Gown AM, Gabbiani G. Intermediate-sized (lo-nm) filaments in human tumors. In: DeLeIlis RA. Advances in immunohis- tochemistry. New York: Masson, 1984:89-109. Ellis GL, Langloss JM, Heffner DK, Hyams VJ. Spindle cell carcinoma of the upper aerodigestive tract: an immunohis- tochemical anaIysisof21 cases. Am JSurg Pathol1987;11:335- 42. Regezi JA, Zarbo RJ, McClatchey KD, Courtney RM, Criss- man JD. Osteosarcomas and chondrosarcomas of the jaws: immunohistochemical correlations. ORAL SURG ORAL MED ORAL PATHOL 1987;64:302-7. Loning TH, Liebsch J, Delling G. Osteosarcomas and Ewing’s sarcomas: comparative immunocytochemical investigation of filamentous proteins and cell membrane determinants. Vir- chows Arch [A] 1985;407:323-36. Roholl PJM, Kleyne J, Elbere H, Van Der Vegt MCD, Albus- Lutter CH, Van Unnik JAM. Characterization of tumor cells in malignant fibrous histiocytomas and other soft tissue tumors in comparison with malignant histiocytes. 1. Immunohis- tochemical study on paraffin sections. J Path01 1985;147:87- 95.

ORAL SURG ORAL MED ORAL PATHOL August 1991

23. McGowan RH. Primary intra-alveolar carcinoma: a difficult diagnosis. Br J Oral Surg 1980;18:259-65.

24. Lucas RB. Pathology of tumors of the oral tissues. 4th ed. New York: Churchill Livingstone, 1984;289-94.

25. Sun T-T, Green H. Keratin filaments of cultured human epidermal cells: formation of intermolecular disulfide bonds during terminal differentiation. J Biol Chem 1978;253: 2053-60.

26. Eichner R, Bonitz P, Sun T-T. Classification of epidermal keratins according to their immunoreactivity, isoelectric point, and mode of expression. J Cell Biol 1984;98:1388-96.

27. Sinclair RA, Burns J, Dunnill MS. Immunoperoxidase stain- ing of formalin-fixed, paraffin-embedded, human renal biop- sies with comparison of the peroxidase-antiperoxidase (PAP) and indirect methods. J Clin Pathol 1981;34:859-65.

28. Dahlin DC, Unni KK. Bone tumors. General aspects and data on 8542 cases. 4th ed. Springfield, III: Charles C Thomas, 1986;286-8.

29. Forteza G, Colmenero B, Lopez-Barea F. Osteogenic sarcoma of the maxilla and mandible. ORAL SURG ORAL MED ORAL PATHOL 1986;62: 179-84.

30. Slootweg PJ, Muller H. Osteosarcoma of the jaw bones: anai- ysis of 18 cases. J Maxillofac Surg 1985;13: 158-66.

31. Batsakis JG. Osteogenic and chondrogenic sarcomas of the jaws. Ann Otol Rhino1 Laryngol 1987;96:474-5.

32. Russ JE, Jesse RH. Management of osteosarcoma of the max- illa and mandible. Am J Surg 1980;140:572-6.

33. Huvos AG. Clinicopathologic spectrum of osteogenic sarcoma: recent observations. Pathol Ann 1979;14:123-44.

34. Garrington GE, Scofield HH, Cornyn J, Hooker SP. Osteo- sarcoma of the jaws: analysis of 56 cases. Cancer 1967;20:377- 91.

35. Rota AN, Smith JL, Jing B-S. Osteosarcoma and parosteal osteogenic sarcoma of the maxilla and mandible: study of 20 cases. Am J Clin Pathol 1970;54:625-36.

36. Zarbo RJ, Regezi JA, Baker SR: Periosteal osteogenic sar- coma of the mandible. ORAL SURC ORAL MED ORAL PATHOL 1984;57:643-7.

37. Chan CW, Kung TM, Ma L. Telangiectatic osteosarcoma of the mandible. Cancer 1986;58:2110-5.

38. Clark JL, Unni KK, Dahlin DC, Devine KD. Osteosarcoma of the jaw. Cancer 1983;51:231 l-6.

39. Lindqvist C, Teppo L, Sane J, Holmstrom T, Wolf J. Osteo- sarcoma of the mandible: analysis of nine cases. J Oral Max- illofac Surg 1986;44:759-64.

41. Bigbee JW, Kosek JC, Eng WF. Effects of primary antiserum dilution on staining of “antigen-rich” tissues with the peroxi- dase antiperoxidase technique. J Histochem Cytochem 1977; 251443-7.

40. Petrusz P. Essential requirements for the validity of immuno- cytochemical staining procedures. J Histochem Cytochem 1983;31:177-9.

42. Grube D. Immunoreactivities of gastrin (G-) cells. II. Non- specific binding of immunoglobulins to G-cells by ionic inter- actions. Histochemistry 1980;66: 149-67.

43. Piris J, Thomas ND. A quantitative study of the influence of fixation on immunoperoxidase staining of rectal mucosal plasma cells. J Clin Pathol 1980;33:361-4.

44. Bosman FT, Lindeman J, Kuiper G, Van der Wal A, Kreunig J. The influence of fixation on immunoperoxidase staining of plasma cells in paraffin sections of intestinal biopsy specimens. Histochemistry 1977;53:57-62.

45. Banks PM. Diagnostic applications of an immunoperoxidase method in hematopathology. J Histochem Cytochem 1979; 2711 192-4.

46. Battifora H, Kopinski M. The influence of protease digestion and duration of fixation on the immunostaining of keratins: a comparison of formalin and ethanol fixation. J Histochem Cy- tochem 1986;34:1095-100.

47. Matthews JB, Mason GI. Influence of decalcifying agents on immunoreactivity of formalin-fixed, paraffin-embedded tissue. Histochem J 1984;16:771-87.

48. Matthews JB. Influence of decalcification on immunohis-

Volume 72 Number 2

Antikeratin immunoreactivity in osteosarcoma of jaw 2 17

49.

50.

51.

52.

53.

54.

tochemical staining of formalin-fixed, paraffin-embedded tis- sue. J Clin Pathol 1982;35:1392-4. Mullink H, Henzen-Logmans SC, Tadema TM, Mol JJ, Meijer CJLM. Influence of fixation and decalcification on the immunohistochemical staining of cell-specific markers in par- affin-embedded human bone biopsies. J Histochem Cytochem 1985;33:1103-9. Mepham BL, Frater W, Mitchell B.S. The use of proteolytic enzymes to improve immunoglobulin staining by the PAP technique. Histochem J 1979;11:345-57. Huang SH. Immunohistochemical demonstration of hepatitis B core and surface antigens in paraffin sections. Lab Invest 1975;33:88-95. Denk H, Radaszkiewicz T, Weirich E. Pronase pretreatment of tissue sections enhances sensitivity of the unlabelled anti- body-enzyme (PAP) technique. J Immune Methods 1977; I 5: 163-7. Curran RC, Gregory J. Demonstration of immunoglobulin in cryostat and paralbn sections of human tonsil by immunoflu- orescence and immunoperoxidase technique: effects of pro- cessing on immunohistochemical performance of tissues and on the use of proteolytic enzymes to unmask antigens in sec- tions. J Clin Pathol 1978;31:974-83. Tseng SCG, Jarvinen MJ, Nelson WC, Huang JW, Wood-

cock-Mitchell J, Sun TT. Correlation of specific keratins with different types of epithelial differentiation: monoclonal anti- body studies. Cell 1982;30:361-72.

55. Woodcock-Mitchell J, Eichner R, Nelson WG, Sun TT. Im- munolocalization of keratin polypeptides in human epidermis using monoclonal antibodies. J Cell Biol 1982;95:580-88.

56. Cooper D, Schermer A, Sun TT. Biology of diseases: classifi- cation of human epithelia and their neoplasms using mono- clonal antibodies to keratins-strategies, applications, and limitations. Lab Invest 1985;52:243-56.

57. Brawer MK, Peehl DM, Stamey TA, Bostwick DG. Keratin immunoreactivity in the benign and neoplastic human pros- tate. Cancer Res 1985;45:3663-7.

58. Gown AM, Vogel AM. Monoclonal antibodies to intermediate filament proteins of human cells: unique in cross reacting an- tibodies. J Cell Biol 1982;95:414-24.

Reprint requesrs: Gary L. Ellis, DDS Department of Oral Pathology Armed Forces Institute of Pathology Washington, DC 20306-6000